You are on page 1of 27

MOSFET

In Fig. 2.1-1, we show an NMOS circuit with its I-V curves and the load line.

Fig. 2.1-1 An NMOS circuit without any AC

In Fig. 2.1-2, an input signal vin is added to VGS . We shall show later Vin
will be amplified. Since the source is signal grounded, we shall call this kind of
amplifiers common source amplifiers. It must be noted that the signal is a small
signal and is also an alternating current (AC) signal. It will become clear that this
kind of amplifiers can hardly amplify large signals.

2-1
Fig 2.1-2 An NMOS circuit with an AC signal

Consider Fig. 2.1-3. As shown in Fig. 2.1-3, the input signal induces an AC
voltage between gate and source. Let us denote this AC voltage by v gs . Throughout
this lecture notes, we shall use small letters to denote AC parameters. Obviously,
v gs  vin . In Fig. 2.1-2, we can see that
Vin  VGS  vin  VGS  v gs
Vout  V DS  v out
I out  I DS  iout

Since the amplitude of v out is larger than that of vin , this circuit functions as an
amplifier. It is important to note the DC voltage V DS still exists. In some sense, we
may say that v out (AC) rides on V DS (DC). But, so far as amplification is concerned,
we are only interested in the AC output voltage signal v out .

2-2
Fig. 2.1-3 The amplification of vin

Suppose a smaller load R L is used, as shown in Fig. 2.1-4. This amplifier will
not function very well now.

Fig. 2.1-4 The behavior of an amplifier with a small load

It will be shown later that a large load is usually desirable and needed. But, this
may drive the transistor out of saturation, as shown in Fig. 2.1-5. Once the transistor
is out of saturation, the output signal will be distorted and the circuit is no longer an
amplifier any more.

2-3
Fig. 2.1-5 A large R L driving the transistor out of saturation

On the other hand, if VGS is not proper, the output signal will also be distorted,
as shown in Fig.2.1-6.

Fig. 2.1-6 An improper VGS causing the distortion of v out

2-4
Section 2.1 The DC Analysis of An NMOS Common
Source Amplifier
In this section, we shall show that the performance of an amplifier can be easily
understood by having a DC analysis of the circuit. That is, we temporarily ignore the
existence of the AC input signal and concentrate our mind on the DC parameters.

Consider Fig. 2.2-1.

Fig. 2.2-1 An amplifier circuit

We shall first see how the value of the load R L affects the performance of the
amplifier. Fig. 2.2-2 shows the three possible cases of R L . As we showed before, if
the load R L is small, the change of VGS will induce only a small change of V DS ,
which is also Vout in our case. On the other hand, for a large R L , even a small
change of VGS will induce a large change of Vout .

2-5
Fig. 2.2-2 I-V curves and different load lines for the circuit in Fig. 2.2-1

Let us assume that the load R L is small. Fig. 2.2-3 shows the relationship
between VGS and Vout . In this case, Vout falls slowly as VGS rises. We added the
AC small signals onto the diagram. It can be seen that v out is quite small. Thus the
amplifier does not function well as an amplifier.

Fig. 2.2-3 Vout vs VGS for small R L

Fig. 2.2-4 shows the situation where R L is large. It can be seen that the sharp
drop of Vout will induce a large output v out and thus a large amplification.

2-6
Fig. 2.2-4 Vout vs VGS for large R L

It should also be seen from Fig. 2.2-5 that in this case, there is a very narrow
region for the proper biasing voltage VGS . In other words, we must be very careful
in selecting the biasing voltage. We shall see later in many experiments that a slight
deviation from a proper VGS may cause great trouble for the amplifier. Thus we may
say that a DC analysis of an amplifier is always important as it gives us information
about how to set the gate bias voltage.

Fig. 2.2-5 The proper selection of the gate biasing voltage

2-7
Section 2.2 The Creation of the AC Current iout in the
Transistor
In the above section, we presented the DC analysis of an amplifier. This explains
why the circuit works as an amplifier. But, it does not give us any idea about the gain
of the amplifier. In other words, it does not let us know how much the input signal
voltage is amplified. To find the gain, an AC analysis is needed.

An amplifier behaves as an amplifier because an AC current is created because


of the input voltage vin , which is also v gs . In the following, we shall first explain
how this AC current, namely iout , is produced. We then proceed to show how this
iout affects v out .

We first point out that the transistor must be in the saturation region. This can be
easily seen in Fig. 2.3-1. If the transistor is in the triode region, the output voltage
would be highly distorted.

Fig. 2.3-1 The case where the transistor is out of saturation

Once the transistor is in the saturation region, we may use Equation (1.2-3)
which we display again as follows:

2-8
1 W 
I out  kn ' (VGS  vin  Vt )
2

2  L 
(2.3-1)
1 W  W  1 W  2
 kn ' (VGS  Vt )  k n ' 
2
(VGS  Vt )vin  k n '  vin
2  L   L  2  L 

In the above equation, the third term is very small because vin is very small and
the first term is a DC term. Thus so far as AC signal is concerned, we have

W 
iout  k n '  (VGS  Vt )vin (2.3-2)
 L

We define

iout W 
gm   kn ' (VGS  Vt ) (2.3-3)
vin L 

Equivalently, we have

iout  g m vin  g m v gs (2.3-4)

Equation (2.3-4) indicates that the input voltage vin creates an AC current iout
linearly because g m is a constant once VGS is fixed. It should be noted that there is
still a DC current I DS . We may say that iout (AC) rides on this I DS (DC) as shown
in Fig. 2.3-2.

Fig. 2.3-2 iout riding on I DS


Section 2.3 The Determination of vout

2-9
In the above section, we showed that the input voltage vin creates an AC current iout
. This current of course would in turn create an AC output voltage v out . We shall
explain its mechanism in this section.

Let us redraw again the amplifier circuit as shown in Fig. 2.4-1.

Fig. 2.4-1 An amplifier circuit

We first show that the iout will affect the output voltage because of the existence
of R L . This can be seen by examining Fig. 2.4-2 which illustrates how iout affects
v out . Note the following:

(1) The input signal vin changes VGS . That is, vin induces an AC voltage
v gs .
(2) This voltage v gs in turn induces a change of current I DS . That is, the AC
voltage v gs induces an AC output current labeled as iout .
(3) As can be seen from Fig. 2.4-2, iout finally changes V DS . That is, it
induces an AC output voltage, labeled as v out .

2-10
Fig. 2.4-2 The inducing of v out by iout

As can be from Fig. 2.4-2, when iout rises, v out falls and when iout falls, v out
rises. Thus, the polarity of v out is opposite to the polarity of iout . From Fig. 2.4-1,
we have

Vout  V DD  I out R L
(2.4-1)
 V DD  ( I DS  iout ) R L

This shows that there is an AC output voltage caused by the existence of R L .


However, we shall not let the AC output voltage simply be equal to - iout R L at this
moment because there is another factor which will affect v out , as we will explain
below.

We usually assume that in the saturation region, the I-V curve of the transistor is
rather flat. In reality, it is not so flat, as illustrated in Fig. 2.4-3.

2-11
Fig. 2.4-3 The meaning of r0

Traditionally, we are familiar with the fact that the change of voltage will cause a
change of current. In this case, it is opposite. If there is a small signal AC current
i ds , a small signal AC voltage v ds will be induced. In other words, the AC current
i ds will induce an AC voltage v ds . Thus, we may define the output impedance r0 as

v ds
r0  (2.4-2)
ids

to illustrate the relationship between i ds and v ds . It must be noted that this r0


exists because of the I-V curves, not because of the load R L . If the slope of the I-V
curve is very flat, r0 will be quite large; otherwise, it will be quite small. A large r0
indicates that a small i ds will include a large v ds .

From the above discussion, we know that v out will be affected by iout , R L and
r0 . In the following, we shall explain how v out is determined.

Let us take a look at the amplifier circuit in Fig. 2.4-1. There are DC voltages in
the circuits. They are VGS and V DD . If they are not proper, the amplifier will not
work. On the other hand, if they are proper, since v out is an AC voltage, it will not
be affected by the existence of these DC voltages. We therefore first short-circuit all
of the DC voltages, namely VGS and VDD . One must note that one terminal of
VDD is connected to R L and another terminal is connected to ground, as shown in
Fig. 2.4-4(a). Once we short-circuit V DD , R L will be connected to ground and the
amplifier circuit becomes that shown in Fig. 2.4-4(b). In Fig 2.4-4(b), there is also an
r0 connected between D and S(ground).

(a) An amplifier circuit with DC voltage (b)The amplifier circuit with DC voltages
short-circuited

2-12
Fig. 2.4-4 The short-circuiting of DC voltages in an amplifier circuit

As we discussed before, a small signal input voltage vin  v gs induces a small


signal current ids . The relationship between v gs and i ds can be found in Equation
(2.3-4). Note they are shown in Equation (2.3-4), v gs and i ds are of the same
polarity. This is illustrated in Fig. 2.4-5. Note that there is no connection between
Node G and Node S as there is no current flowing into the gate. That is, there is a
voltage between G and S, but no current from G..

Fig. 2.4-5 Small signal equivalent circuit of a transistor

Because of the short-circuiting of DC voltages, R L is across D and S as shown


in Fig. 2.4.4(b). So is r0 . The entire small signal equivalent circuit of the
amplifier is shown in Fig.2.4-6(b).

(a) An amplifier circuit (b) The small signal equivalent circuit (c) The small signal equivalent
circuit further simplified

Fig. 2.4-6 The small signal equivalent circuit of an amplifier circuit

2-13
From Fig. 2.4-6(c), we can see that the current will flow through the parallel
connection of R L and r0 , which is expressed as Z  R L // r0 . The value of Z can be
found as follows:

1 1 1
 
Z RL r0

R L r0
Z
RL  r0

v out   g m vin ( R L // r0 )
R L r0 (2.4-3)
  g m vin
R L  r0

Note that the polarity of v out is opposite to that of vin , as expected by


examining Fig. 2.4-2. The meaning of the opposition of polarity is illustrated in Fig.
2.4-7.

Fig. 2.4-7 The meaning of polarity of AC signals

Let us note that r0 is usually much larger than R L because in reality, the I-V
curves of a transistor are rather flat and a flat I-V curve produces a large r0 . On the
other hand, as we pointed out before, R L cannot be too large because a large R L
will drive the transistor into the non-saturation region. When a large resistor is in
parallel with another resistor, it can be ignored. Thus, r0 is usually ignored and we
have

v out   g m R L vin (2.4-4)

2-14
We often denote the gain of the amplifier by Av . Then

v out
Av    g m RL (2.4-5)
vin

That a large R L will produce a better amplifier can also be seen by the DC
analysis. The reader should go back to Section 2.2 again. Take a look at the figures
from Fig. 2.2-2 to Fig. 2.2-5. As seen in these figures, we can see that a large R L
will produce a sharp input-output relationship and thus a high gain.

In the above sections, we only talked about NMOS amplifiers. The same
discussion can be used to explain how a PMOS amplifier works. A typical PMOS
amplifier is shown in Fig. 2.4-8.

Fig. 2.4-8 A PMOS amplifier


Section 2.4 Experiments of the Design of NMOS
Transistors
To design an NMOS amplifier, we must pay attention to the operating point. Fig.
2.5-1 shows how the operating point is determined. It is essentially determined by
selecting an appropriate R L and an appropriate VGS .

2-15
Fig. 2.5-1 The determination of operating points

In the following, we shall show a set of experiments to demonstrate how


different parameters, including VGS and R L , will affect the performance of the
amplifier.

Experiment 2.5-1 An Appropriate Operating Point

The purpose of this experiment is to demonstrate that an appropriate operating


point can be determined. The circuit is shown in Fig. 2.5-2, the SPICE program is in
Table 2.5-1 and the operating point is shown Fig. 2.5-3. The input and output signals
are shown in Fig. 2.5-4. The gain of this amplifier was found to be 20.

Fig. 2.5-2 The amplifier circuit with R1  100 K for Experiment 2.5-1

2-16
Table 2.5-1 The program for Experiment 2.5-1
EX2-1
.protect
.lib 'D:\model\tsmc\MIXED035\mm0355v.l' TT
.unprotect
.op
.options nomod post

VDD 1 0 3.3v
R1 1 2 100k
V2 2 0 0v

.param W1=5u
M1 2 3 0 0
+nch L=0.35u W='W1' m=1
+AD='0.95u*W1' PD='2*(0.95u+W1)'
+AS='0.95u*W1' PS='2*(0.95u+W1)'

*Fig2.5-3*
VGS_1 3 0 0.65v
.DC V2 0 3.3v 0.1v
.PROBE I(R1) I(M1)

*Fig2.5-4*
*VGS_2 3 4 0.65v
*Vin 4 0 sin(0v 0.01v 10Meg)
*.tran 0.1ns 600ns

.end

2-17
IDS

Load Line I-V Curve

VDS
Fig. 2.5-3 The operating points of the circuit in Fig. 2.5-2

vout

vin

Fig. 2.5-4 The gain of the amplifier in Experiment 2.5-1

2-18
Experiment 2.5-2 The Operating Point with the Same VGS, but a Smaller Load

In this experiment, we used a smaller load as shown in Fig. 2.5-5. The program
is in Table 2.5-2, the operating point is shown in Fig. 2.5-6 and the input and output
voltages are shown in Fig. 2.5-7. As can be seen in Fig. 2.5-7, the gain is now
smaller.

Fig. 2.5-5 The amplifier circuit with R1  16k for Experiment 2.5-2

Table 2.5-2 Program for Experiment 2.5-2


EX2-2
.protect
.lib 'D:\model\tsmc\MIXED035\mm0355v.l' TT
.unprotect
.op
.options nomod post

VDD 1 0 3.3v
R1 1 2 10k
V2 2 0 0v

.param W1=5u
M1 2 3 0 0
+nch L=0.35u W='W1' m=1
+AD='0.95u*W1' PD='2*(0.95u+W1)'
+AS='0.95u*W1' PS='2*(0.95u+W1)'

2-19
*Fig2.5-3*
VGS_1 3 0 0.65v
.DC V2 0 3.3v 0.1v
.PROBE I(R1) I(M1)

*Fig2.5-4*
*VGS_2 3 4 0.65v
*Vin 4 0 sin(0v 0.01v 10Meg)
*.tran 0.1ns 600ns

.end

IDS

Load Line

I-V Curve

VDS

Fig. 2.5-6 The operating point of the amplifier of Experiment 2.5-2

2-20
vout

vin

Fig. 2.5-7 The gain of the amplifier in Experiment 2.5-2

Experiment 2.5-3 An Amplifier with a Higher VGS

If VGS is too high, the transistor may be outside of saturation and the output
signal will be distorted. The purpose of this experiment is to demonstrate this point.
We raised VGS from 0.65V to 0.72V. As can be seen, the transistor is indeed out of
saturation and the output signal is indeed distorted. The testing is shown below.

2-21
Fig.2.5-8 The amplifier circuit with VGS  0.72 for Experiment 2.5-3

Table 2.5-3 Program for Experiment 2.5-3


EX2-3
.protect
.lib 'D:\model\tsmc\MIXED035\mm0355v.l' TT
.unprotect
.op
.options nomod post

VDD 1 0 3.3v
R1 1 2 100k
V2 2 0 0v

.param W1=5u
M1 2 3 0 0
+nch L=0.35u W='W1' m=1
+AD='0.95u*W1' PD='2*(0.95u+W1)'
+AS='0.95u*W1' PS='2*(0.95u+W1)'

*Fig2.5-3*
VGS_1 3 0 0.72v

2-22
.DC V2 0 3.3v 0.1v
.PROBE I(R1) I(M1)

*Fig2.5-4*
*VGS_2 3 4 0.72v
*Vin 4 0 sin(0v 0.01v 10Meg)
*.tran 0.1ns 600ns

.end

IDS
I-V Curve

Load Line

VDS

Fig. 2.5-9 The operating points of the amplifier of Experiment 2.5-3

2-23
Fig. 2.5-10 The distortion of output signals of the amplifier in Experiment 2.5-3

Section 2.5 Some Guidelines for Designing MOSFET


Amplifiers
By now, the reader may have some feeling about the designing of an NMOS, or
PMOS, amplifier. The important thing is to determine an appropriate operating point.
Suppose that we start with the situation as shown Fig. 2.6-1.

Fig. 2.6-1 The determination of the operating point of NMOS

2-24
We may either increase VGS which will result in the situation shown in Fig. 2.6-
2(a) or increase R L which will result in the situation in Fig. 2.6-2(b).

Fig. 2.6-2 The adjusting of parameters of an amplifier to obtain appropriate operating


points

We must note that an increase of VGS will result in a larger I DS . It is not very
ideal to have a large I DS as this will increase the power consumption.

In general, we would like to have a small I DS which corresponds to a low VGS .


Note that VGS has to be higher than Vt . Usually, VGS is not significantly higher
than Vt . If Vt is around 0.55V, it is appropriate to set VGS to be 0.65V.

` Let us consider Fig. 2.6-3. In this case, assume that we have a rather large VGS .
This will not allow R L to be large because a large R L will drive the transistor into
the non-saturation region. Thus we must use a smaller load if VGS is large, which is
not ideal.

Fig. 2.6-3 The problem of using a large VGS

2-25
The reader is also encouraged to take a look at Fig. 2.2-5 to see why VGS cannot
be too large.

In any case, this kind of amplifiers suffer from one drawback. We like R L to be
large. Yet, R L cannot be too large because a large R L will often drive the transistor
out of saturation. In the next chapter, we shall introduce CMOS circuits which will
avoid this problem

Section 2.7 Exercise 2

1. Explain why the following is an amplifier in Fig. 2.7-1.

Fig. 2.7-1 The figure for Problem 1 in Exercise 2

2. In the above circuit, explain why you cannot use a very small load?
3. In the above circuit, explain why you cannot use a very large load?
4. Consider the Vout vs VGS curve.
(a) Do you want a curve which is rather sharp or a curve which is rather flat?
Why?

2-26
(b) How do you achieve a sharp Vout vs VGS curve?
5. Derive the following formula: iout  g m v gs .
6. Explain why we need to short circuit all constant voltage sources in the small
signal equivalent circuit of an amplifier.
7. Consider Fig. 2.7-2.

Fig. 2.7-2 The figures for Problem 7 of Exercise 2

(a) Explain the meaning of r0 .


(b) Explain why r0 can be usually ignored.
(c) Explain how this equivalent circuit is obtained.
(d) Derive the following formula: vout   g m R L vin .
(e) Explain the meaning of the negative sign.
(8) In general, what is the value of VGS ?

2-27